Oncogene (2006) 25, 2444–2451
& 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc
ORIGINAL ARTICLE
Induction of tetraploidy through loss of p53 and upregulation of Plk1 by human papillomavirus type-16 E6 A Incassati1, D Patel2 and DJ McCance2,3 1 Department of Biomedical Genetics, University of Rochester, Rochester, NY, USA; 2Department of Microbiology and Immunology, University of Rochester, Rochester, NY, USA and 3James P. Wilmot Cancer Center, University of Rochester, Rochester, NY, USA
Cancer cells are insensitive to many signals that inhibit growth of untransformed cells. Here, we show that primary human epithelial cells expressing human papillomavirus (HPV) type-16 E6/E7 bypass arrest caused by the DNA-damaging drug adriamycin and become tetraploid. To determine the contribution of E6 in the context of E7 to the resistance of arrest and induction of tetraploidy, we used an E6 mutant unable to degrade p53 or RNAi targeting p53 for knockdown. The E6 mutant fails to generate tetraploidy; however, the presence of E7 is sufficient to bypass arrest while the p53 RNAi permits both arrest insensitivity and tetraploidy. We published previously that polo-like kinase 1 (Plk1) is upregulated in E6/E7-expressing cells. We observe here that abnormal expression of Plk1 protein correlates with tetraploidy. Using the p53 binding-defective mutant of E6 and p53 RNAi, we show that p53 represses Plk1, suggesting that loss of p53 results in tetraploidy through upregulation of Plk1. Consistent with this hypothesis, overexpression of Plk1 in cells generates tetraploidy but does not confer resistance to arrest. These results support a model for transformation caused by HPV-16 where bypass of arrest and tetraploidy are separable consequences of p53 loss with Plk1 required only for the latter effect. Oncogene (2006) 25, 2444–2451. doi:10.1038/sj.onc.1209276; published online 12 December 2005 Keywords: HPV-16; E6 and E7; tetraploidy; Plk1
Introduction The process of carcinogenesis renders a cell unresponsive to the checks and balances acting throughout a normal cell cycle to preserve genome integrity. As a result, cancer cells frequently display genetic instability in the form of aneuploidy. The mechanisms underlying Correspondence: Dr DJ McCance, Department of Microbiology and Immunology, Box 672, University of Rochester, Rochester, NY 14642, USA. E-mail:
[email protected] Received 10 August 2005; revised 28 September 2005; accepted 19 October 2005; published online 12 December 2005
the manifestation of the aneuploid state are elusive; however evidence exists, for example from the studies of Barrett’s esophagus, suggesting a premalignant cell subsists as a tetraploid intermediate before the onset of aneuploidy and full transformation (Galipeau et al., 1996; Storchova and Pellman, 2004). In light of this observation, attention has focused on the contribution of dysfunctional mitotic checkpoint proteins to the development of aneuploidy. Loss of function mutations in BUB1 and BUB1B genes result in mitotic failure and subsequent aneuploidy (Cahill et al., 1998; Hanks et al., 2004; Michel et al., 2004). Conversely, overexpression of the mitotic kinases Aurora-A or Polo-like kinase-1 (Plk1) causes centrosome amplifications and tetraploidy resulting from multinucleation (Meraldi et al., 2002). These proteins are mutated or upregulated in various human tumors and cancer cells (Nigg, 2001), thus supporting the notion that mis-expression of kinases that normally function to ensure faithful mitotic progression, may contribute to the evolution of disease through the generation of chromosomal instabilities. Cervical squamous cell carcinoma is a cancer type in which cells are aneuploid. Cervical carcinogenesis is tightly correlated with infection by the high-risk human papillomaviruses (HPVs) and particularly, HPV type-16 is present in nearly half of cervical cancer specimens examined (Burd, 2003). The transforming ability of the virus is attributed to the unregulated expression of two viral oncoproteins, E6 and E7 (E6/E7), upon viral integration into the host cell genome. E7 disrupts the function of the retinoblastoma (Rb) family of cell cycle regulators (Dyson et al., 1989). E6 binds and degrades the tumor suppressor p53 (Scheffner et al., 1990; Werness et al., 1990). The functions of each viral protein have been examined thoroughly in isolation, however, these analyses do not accurately represent the in vivo scenario in cases of cervical cancer, where both E6 and E7 are always present (zur Hausen, 2000). The relevance of this issue is illustrated in a tissue culture model using primary human foreskin keratinocytes (HFKs), the host cell of the virus. Prolonged coexpression of E6/E7 in HFKs efficiently immortalizes these cells in defined media, while expression of either individual oncoprotein does not (Hawley-Nelson et al., 1989; Munger et al., 1989; McMurray and McCance, 2004). Coexpression of E6/E7 in HFKs effectively
Upregulation of Plk1 by HPV-16 E6 results in tetraploidy A Incassati et al
2445
recapitulates cervical carcinogenesis triggered by HPV and can be used to model the initial stages of epithelial premalignancies in a tissue culture system. We and others have shown that E6/E7 expression in HFKs causes bypass of cell cycle arrest induced by spindle checkpoint activation upon nocodazole treatment. E6/E7-expressing cells exposed to nocodazole consequently became tetraploid (Thompson et al., 1997; Thomas and Laimins, 1998; Patel et al., 2004). Bypass of arrest and generation of tetraploidy correlated with the aberrant upregulation of Plk1 by E6/E7 in a manner dependent upon inhibition of p53 or Rb by E6 or E7, respectively (Patel et al., 2004). This prior investigation implicated a role for Plk1 dysregulation in the phenotypes caused by E6/E7 in primary keratinocytes. Previous studies regarding Plk1 function in human cells have focused predominantly on Plk1 function in cancer cell lines, however, the role of Plk1 in untransformed human cells has remained unclear (Spankuch-Schmitt et al., 2002; Guan et al., 2005). In this report, we characterize the contribution of Plk1, to the bypass of an arrest signal caused by the DNA-damaging agent adriamycin and the induction of tetraploidy in primary human epithelial cells. We provide evidence that Plk1 functions downstream of p53 degradation by E6 to produce tetraploidy but does not contribute to the bypass of arrest by E6/E7. Our results suggest that the bypass of DNA-damage-induced arrest and tetraploidy are two separable events in the progression of cancer. Results Cells expressing HPV-16 E6/E7 bypass DNA damageinduced arrest and become tetraploid HFKs were infected with pBabe retrovirus encoding polycistronic E6 and E7 (E6/E7) from the high-risk HPV type-16 to generate stable lines expressing E6/E7 (pBP.E6/E7) or vector alone (pBP). The cells were treated with adriamycin, a drug that generates DNA double-strand breaks and arrests HFKs in G1 and G2 phases (Flatt et al., 1998). After 24 h of adriamycin exposure, cells were pulsed with bromodeoxyuridine (BrdU) and nuclei were harvested for fluorescenceactivated cell sorting (FACS) analysis. In the presence of adriamycin, the majority of pBP cells arrested as only 4.4% of the cells incorporated BrdU (Figure 1a, see %S). In contrast, 24.9% of pBP.E6/E7 cells continued cycling in the presence of adriamycin (Figure 1a, see %S). We obtained similar results using etoposide, a topoisomerase II inhibitor, which arrests cells in G2 (data not shown). We further corroborated this arrest circumvention by E6/E7 in IMR-90 fibroblasts, an additional nontransformed human cell type (data not shown). HFKs expressing E6/E7 from the low-risk HPV types-6 and -11, which are nononcogenic, arrested when treated with adriamycin, with only 7.4 and 8.6% of cells, respectively, continuing to incorporate BrdU (Figure 1a). RT-PCR analysis confirmed the expression of E6/E7 mRNA in the high- and low-risk HPV subtypes (Figure 1b).
Figure 1 Bypass of adriamycin-induced arrest and generation of tetraploidy in cells expressing HPV-16 E6/E7. (a) HFKs expressing E6/E7 from HPV type-16, -6, or -11 or vector were treated with adriamycin, pulsed with BrdU and harvested for FACS analysis as described in Materials and methods. DNA content as measured by PI stain is plotted on the X-axis. BrdU incorporation is plotted on the Y-axis of the dot plots and total counts are plotted on the Yaxis of the histograms. Arrows indicate the nuclei populations with 2N or 4N DNA content. (b) RT-PCR analysis of HPV-16, -6, or 11 E6/E7 in HFKs described in (a) PCR products run at 750 bp for all transcripts.
Interestingly, within the total population of pBP.E6/ E7 HFKs, adriamycin treatment revealed a large percentage of cells that were tetraploid (Figure 1a, see %>4N). These results are not explained by multinucleation, as individual nuclei were isolated for FACS. Instead the tetraploid genome resided within single nuclei. A majority of tetraploid cells appeared to be part of the subpopulation of cells, which incorporated BrdU in the presence of adriamycin, suggesting that these two phenotypes are interconnected. These experiments show that expression of E6/E7 from the oncogenic HPV type16 results in evasion of DNA damage arrest signals and the creation of a tetraploid subpopulation of cells. We next focused on dissecting the functions of E6/E7 responsible for the observed phenotypes. Bypass of adriamycin-induced arrest and generation of tetraploidy by E6/E7 requires loss of p53 The ability of E6 to degrade p53 has been shown to contribute to the bypass of DNA damage and spindleinhibiting arrest signals by E6/E7 (Foster et al., 1994; Thompson et al., 1997; Patel et al., 2004). Protein levels of p53 are significantly lower in pBP.E6/E7 HFKs than in cells expressing pBP or E6/E7 from the low-risk HPV types, especially upon adriamycin treatment (Figure 2a). Oncogene
Upregulation of Plk1 by HPV-16 E6 results in tetraploidy A Incassati et al
2446
To investigate whether E6 degradation of p53 played a causal role in the arrest insensitivity and tetraploidy formation observed in pBP.E6/E7 cells, a mutant of E6 with a deletion of residues 123–127, which renders E6 unable to bind or degrade p53 was utilized (Foster et al., 1994). Western blot analysis revealed an equivalent amount of E7 protein expressed in pBP.E6/E7 cells and cells expressing the E6 mutant (pBP.E6D123-127/E7), while in the context of wild-type E7 (Figure 2b). Cells were then treated with adriamycin and analysed for p53 protein quantity. More abundant levels of p53 were observed in the E6 mutant cells than in pBP controls (Figure 2b). E7 has been shown to stabilize p53 protein, albeit in an inactive state (Eichten et al., 2002). Therefore, a more accurate assessment of functional p53 levels in this scenario is attained by Western blotting for the active form of p53 phosphorylated on serine 15. Cells expressing the E6 mutant had comparable amounts of phosphorylated p53 as pBP cells upon adriamycin treatment (Figure 2b). Significantly, cells expressing the E6 mutant in the context of E7 did not become tetraploid, suggesting that p53 degradation by E6 is necessary to generate tetraploidy. However, these cells were insensitive to adriamycin-induced arrest
Figure 2 Generation of tetraploidy, but not bypass of arrest by HPV-16 E6/E7-expressing cells, requires p53 degradation. (a) HFKs expressing HPV-16, -6, or -11 E6/E7 or vector were adriamycin-treated and harvested for p53 western blot. A nonspecific band at 110 kDa recognized by the anti-Plk1 antibody (NS) is used for loading control. (b) HFKs expressing E6/E7 or E6.D123-127/E7 were treated with adriamycin and harvested for Western blotting analysis of p53 or p53 phosphorylated on ser-15. Western blot for E7 was done on untreated pBP, E6/E7 or E6.D123-127/E7 HFKs. (c) HFKs described in (b) were treated with adriamycin, pulsed with BrdU and harvested for FACS analysis. Oncogene
(Figure 2c). The presence of E7 is likely responsible for the bypass of arrest seen here, as E7 expression has been shown to permit resistance to various arrest signals (Demers et al., 1996). In order to avoid any contribution E7 may impart to the initiation of tetraploidy or bypass of arrest, we decided to address directly whether loss of p53 is sufficient to circumvent arrest and induce tetraploidy by producing HFKs with stable knockdown of p53 protein (pSuper.retro (pSR).p53i) using the pSuper retroviral system to stably express a sequence targeting p53 for degradation (Brummelkamp et al., 2002). The knockdown cells had lower levels of p53 protein compared to cells expressing a scrambled RNAi control (pSR.scrambled) in the presence or absence of adriamycin (Figure 3a). Decreased p53 protein levels resulted in the circumvention of adriamycin-induced arrest and induction of tetraploidy as compared to control cells (Figure 3b). In addition, direct comparison of the pSR.p53i cells with the E6/E7-expressing cells revealed that an equivalent percentage of each cell population continued to bypass adriamycin-induced arrest and established tetraploidy, while control pBP or pSR.scrambled cells arrested without becoming tetraploid (Table 1). The RNAi experiments further supported the hypothesis that p53 degradation by E6 was sufficient to allow bypass of the DNA damage response and the generation of tetraploid cells, suggesting that both of these phenotypes are regulated through p53. Nonetheless, the mutational analysis of E6 showed that the ability of E6/E7-expressing cells to resist arrest may actually occur through the cooperation of both oncoproteins, while the induction of tetraploidy is mediated largely by E6. These results collectively suggest that in the process of HPV-mediated transformation, the two phenotypes may actually be separable. We next sought to determine whether these phenotypes resulting from expression of E6 and E7 were the result of distinct mechanisms.
Figure 3 Targeted loss of p53 is sufficient to bypass adriamycininduced arrest and generate tetraploidy. (a) pSR.scrambled control or pSR.p53i HFKs were treated with adriamycin and harvested for Western blotting analysis of p53 or a nonspecific loading control (NS) as described in Figure 2. (b) Cells treated above were pulsed with BrdU and harvested for FACS analysis.
Upregulation of Plk1 by HPV-16 E6 results in tetraploidy A Incassati et al
2447
Plk1 protein expression correlates with tetraploidy and is negatively regulated by p53 Previously, we had shown that E6/E7-expressing HFKs have increased levels of polo-like kinase 1 (Plk1) protein (Patel et al., 2004). Plk1 has furthermore been shown to be inactivated upon DNA damage (Smits et al., 2000). Recently, it was demonstrated that Plk1 is repressed by p53 upon an apoptotic signal (Kho et al., 2004). For these reasons, Plk1 appeared to be an attractive candidate for a protein that would contribute to the phenotypes caused by expression of E6/E7. When pBP HFKs were exposed to adriamycin, Plk1 protein levels decreased to undetectable amounts (Figure 4a). In matched pBP.E6/E7 cells, however, Plk1 levels were upregulated upon adriamycin treatment (Figure 4a). We noticed that Plk1 protein levels in E6/E7-expressing cells fluctuated slightly between experimental replicates (data not shown). This small variation can be attributed to the use of different pools of primary keratinocytes, each with varying genetic backgrounds, for each experiment. Despite the minor fluctuations observed, the levels of Plk1 in E6/E7-expressing cells upon adriamycin treat-
Table 1 HFKs expressing E6/E7 or p53 RNAi bypass adriamycininduced arrest and become tetraploid to a similar extent Sample pBP E6/E7 pSR.scrambled pSR.p53i pBP E6/E7 pSR.scrambled pSR.p53i
7 Adriamycin
%S
%>4N
+ + + +
32.8 49.7 32.2 45.1 2.1 17.8 6.1 21.4
11.1 30.6 10.9 26.8 16.3 37.4 5.9 34.1
Cells expressing pBP, pBP-E6/E7, pSR.scrambled RNAi control, or pSR.p53i were treated with adriamycin, pulsed with Brdu, and harvested for FACS. This data set represents one experiment of three.
ment remain grossly elevated as compared to pBP controls, indicating a disruption of Plk1 regulation by E6/E7 in response to adriamycin-induced arrest. Plk1 was also upregulated in IMR-90 fibroblasts expressing E6/E7 as compared to pBP cells (data not shown). Additionally, the level of Plk1 protein in HFKs expressing the nononcogenic HPV types-6 and -11 decreased upon adriamycin exposure in a manner as seen in control pBP cells (Figure 4a). We next examined whether Plk1 was upregulated in cells expressing the mutant E6 unable to degrade p53. Plk1 levels were elevated somewhat in untreated HFKs expressing the E6 mutant in the context of E7 as compared to pBP controls. Yet, upon adriamycin treatment Plk1 protein levels were repressed in E6 mutant cells, while levels of Plk1 observed in E6/E7expressing cells had increased (Figure 4b). This result suggested that the ability of E6 to degrade p53 was necessary to upregulate Plk1. To more specifically address whether loss of p53 was sufficient to increase Plk1 protein upon adriamycin treatment, the pSR.p53i cells were treated and examined for Plk1 protein levels. Plk1 was elevated in these cells upon adriamycin treatment to the same extent as cells expressing E6/E7, as compared to cells expressing pBP empty vector or the pSR.scrambled RNAi control (Figure 4c). We were curious to examine whether the kinetics of Plk1 protein regulation in pBP.E6/E7 cells upon adriamycin treatment mirrored that of the pSR.p53i cells. To examine this, Western blotting was performed on pBP, pBP.E6/ E7 and pSR.p53i cells at 6, 12, and 24 h post-treatment with adriamycin (Figure 4d). As rapidly as 6 h postexposure to adriamycin, p53 protein was upregulated in pBP HFKs. Subsequently, a decrease and then disappearance of Plk1 was observed by 24 h. In contrast, both pBP.E6/E7 and pSR.p53i exhibited similar trends of Plk1 regulation. Both cell lines showed a mild induction of p53 upon adriamycin treatment and increased Plk1
Figure 4 Upregulation of Plk protein levels upon adriamycin treatment correlates with bypass of adriamycin-induced arrest and tetraploidy. (a) HFKs expressing HPV-16, -6, or -11 E6/E7 were treated with adriamcyin and harvested for Western blotting analysis for Plk1 or a nonspecific loading control (NS) detected by the anti-Plk1 antibody. (b) Cells expressing E6/E7 or E6.D123-127/E7 were treated with adriamcyin and harvested for Plk1 Western blot. (c) pBP, E6/E7, pSR.scrambled control or pSR.p53i cells were treated with adriamycin and harvested for Plk1 Western blot. (d) pBP, E6/E7-expressing HFKs or pSR.p53i cells were treated for 6, 12, or 24 h with adriamycin and harvested at each time point for p53 and Plk1 Western blots. Oncogene
Upregulation of Plk1 by HPV-16 E6 results in tetraploidy A Incassati et al
2448
protein by 12 h (Figure 4d). These experiments imply that the aberrant upregulation of Plk1 observed in primary epithelial cells expressing E6/E7 after treatment with adriamycin is due to the loss of p53. The regulation of Plk1 by p53 had been investigated thus far by using stable cell lines. We next examined whether Plk1 was regulated by p53 in an acute fashion as demonstrated by transient transfection experiments. Exogenous expression of p53 in HFKs caused a decrease of Plk1 protein by 48 h post-transfection (Figure 5a). Conversely, transient expression of a siRNA targeting p53 was also utilized. To address possible off-target effects resulting from the RNAi molecule used to generate the pSR.p53i stable lines, a distinct p53 siRNA sequence was employed for the transient assays. Transient knockdown of p53 by siRNA resulted in increased Plk1 protein levels (Figure 5b). Plk1 has been shown to be transcriptionally repressed by p53 in colon carcinoma and postcrisis fibroblast cells (Jackson et al., 2005). Using a Plk1 promoter luciferase construct to examine the transcriptional regulation of Plk1 by p53 in primary human epithelial cells, exogenous expression of p53 repressed the Plk1 promoter 10fold compared to basal activation in cells transfected with empty vector (Figure 5c). Additionally, knocking
Figure 5 Plk is negatively regulated by p53. (a) HFKs transfected with pcDNA or pcDNA-p53 were harvested after 48 h for Western blotting analysis of p53, Plk1 or a nonspecific band recognized by the anti-Plk1 antibody (NS). Numbers below bands refer to relative ratios of Plk1 band intensities determined by densitometer analysis. (b) HFKs transfected with pBS-U6 p53 siRNA or pBS-U6 luciferase siRNA were harvested after 48 h for p53 and Plk1 Western blots. Numbers below bands refer to relative ratios of p53 band intensities quantified by densitometer. (c) HFKs transfected with 250 ng pGL2-Plk1 luciferase reporter and either 750 ng pcDNA or pcDNA-p53 were lysed after 48 h and luciferase activity was measured. (d) HFKs were transfected with 250 ng pGL2-Plk1 luciferase reporter and either 750 ng pSG5-E6/E7 or 750 ng pBSU6 p53 siRNA. For negative controls, cells were transfected with reporter and 750 ng pSG5 or 750 ng pBS-U6 control (Akt3 siRNA) as a nonrelevant target. Luciferase activity for each control was made equal to one and luciferase activity from cells transfected with E6/E7 or p53 siRNA was normalized to each appropriate control. Cells were treated with adriamcyin and lysed after 48 h for luciferase assay. Oncogene
down p53 activated the Plk1 promoter to nearly the same extent as expression of E6/E7, when compared to the basal activity of the Plk1 promoter transfected with pSG5 empty vector or a control siRNA (Figure 5d). Activation of the Plk1 promoter by either E6/E7 or p53 siRNA transfection was observed in cells treated or untreated with adriamycin (Figure 5d). The above experiments show that p53 negatively regulates Plk1 transcription and protein in the presence or absence of adriamycin in HFKs and that this contributes to the abnormal upregulation of Plk1 by E6/E7. Plk1 overexpression induces tetraploidy, but is not sufficient to bypass arrest To examine directly whether upregulation of Plk1 in HFKs was a contributing factor to the bypass of adriamycin-induced arrest or generation of tetraploidy in E6/E7-expressing cells, we created HFKs stably overexpressing Plk1 (Figure 6a). Upon adriamycin treatment, the level of Plk1 in these cells decreased, presumably due to the functional p53 and reduction in the levels of endogenous Plk1 (Figure 6a). However, exogenous Plk1 remained clearly detectable upon
Figure 6 Overexpression of Plk1 is sufficient to induce tetraploidy but not bypass of arrest. (a) HFKs expressing E6/E7 or Plk1 were treated with adriamcyin and harvested for Plk1 Western blot or a nonspecific band recognized by the anti-Plk1 antibody (NS). (b) Cells treated above were pulsed with BrdU and harvested for FACS analysis. (c) Schematic of E6 and E7 functions contributing to the bypass of DNA damage arrest and induction of tetraploidy.
Upregulation of Plk1 by HPV-16 E6 results in tetraploidy A Incassati et al
2449
treatment. While overexpression of Plk1 effectively induced an equivalent tetraploid cell population as E6/ E7-expressing cells in the absence of drug, the Plk1 overexpressing cells remained sensitive to DNA damage arrest (Figure 6b). This result corroborates the data obtained from the E6 mutational analysis where expression of the E6 mutant in the context of E7 confers resistance to arrest but does not induce tetraploidy or upregulate Plk1 (Figures 2a and 3b). These data imply that although bypass of arrest and generation of tetraploidy are both phenotypes regulated by p53, they are in fact two separate events. Figure 6c shows a schematic of our proposed hypothesis. Disruption of p53 by E6 causes both bypass of adriamycininduced arrest and tetraploidy, while E7 contributes to the bypass of arrest. Although, Plk1 may be a downstream target of p53 in the prevention of tetraploidy, it does not partake in the p53 response to DNA damage, supporting a branching of the p53 effector pathway. Discussion In this study, we investigated the role of Plk1 as a consequence of p53 disruption in the bypass of adriamycin-induced arrest and the generation of tetraploidy in primary human epithelial cells expressing HPV-16 E6/E7. Expression of either E6 or E7 has been shown to confer resistance to multiple arrest signals, including actinomycin-D, TGF-b, and nocodazole (Foster et al., 1994; Demers et al., 1996; Thomas and Laimins, 1998; Patel et al., 2004). The properties of E6 and E7 most important for circumventing arrest were degradation of p53 and abrogation of Rb function, respectively (Foster et al., 1994; Demers et al., 1996; Thompson et al., 1997; Patel et al., 2004). Targeted loss of p53 has been shown to result in both the bypass of DNA damage-induced arrest as well as generation of tetraploidy (Cross et al., 1995; Lanni and Jacks, 1998). Using a mutant of E6 unable to degrade p53, we show that E6 inhibition of p53 is required for tetraploidy induction, while expression of E7 in these cells was insufficient to generate tetraploidy. These results are in contrast to prior work examining the ability of E7 to induce tetrasomy in HFKs (Southern et al., 2004). The tetrasomy observed by Southern et al. (2004) was scored by visualizing chromosome markers using immunoflouresence. This analysis method may be more sensitive than FACS, and therefore may have detected low levels of tetraploidy that we were unable to detect here, although the levels of tetrasomy in E7 expressing cells was in the range observed here in control cells containing an empty vector. Another factor contributing to the discrepancy involves the level of E7 expression obtained by the different retroviral vector systems employed. The E7 protein expression may have been higher in this previous analysis, and therefore resulted in dosedependent effects of E7, which were not observed here. In our experiments, expression of E7 in the context of the E6 mutant did not cause tetraploidy over back-
ground levels. These cells, however, continued to resist arrest due to the presence of E7, which masked the effects of sustained p53 regulation. Directly knocking down p53 by RNAi was sufficient to allow bypass of arrest as well as induction of tetraploidy to the same extent as E6/E7-expressing cells, suggesting that E6 degradation of p53 is required for the generation of tetraploidy. To determine a possible mechanism by which E6 may be mediating tetraploidy through loss of p53, we investigated the role of Plk1 which we had shown previously to be cooperatively upregulated by E6/E7 in primary human keratinocytes (Patel et al., 2004). Plk1 is an important modulator of many aspects of mitosis (Nigg, 2001). Overexpression of Plk1 causes inappropriate mitotic events such as disrupted cytokinesis resulting in abnormalities like multinucleation in cancer cell lines (Mundt et al., 1997; Meraldi et al., 2002). However, Plk1 overexpression experiments have not previously been carried out in primary human cells. Generation of HFKs stably expressing high levels of Plk1 protein resulted in tetraploidy within individual nuclei to the same extent as cells expressing E6/E7, suggesting that upregulation of Plk1 by E6/E7 may contribute to the induction of tetraploidy. Abundance of Plk1 has been linked with poor prognosis in human cancers including ovarian and head and neck cancers (Knecht et al., 1999; Weichert et al., 2004, 2005). Our data support a role for inappropriate Plk1 expression early in malignant conversion caused by HPV, contributing to the formation of tetraploid cells potentially through cytokinesis failure, which may eventually lead to the development of aneuploid cells. Plk1 protein levels in cells with knockdown of p53 were equivalent to cells expressing E6/E7. Subsequent analysis showed that Plk1 was negatively regulated by p53, in agreement with recently published results describing the inhibition of Plk1 by p53 upon an apoptotic signal (Kho et al., 2004). Our data suggested that E6/E7 may be upregulating Plk1 through loss of p53 resulting directly in bypass of adriamycin-induced arrest and generation of tetraploidy. However, experiments utilizing HFKs exogenously expressing Plk1 as well as the E6 mutational analysis revealed that p53 degradation and upregulation of Plk1 accounted for induction of tetraploidy but not bypass of arrest. The mutational analysis also showed that E7 had little effect on total Plk1 protein in HFKs treated with adriamcyin. E7 has been shown to regulate Plk1 under various conditions. We have shown previously that E7 is capable of regulating Plk1 upon mitotic spindle inhibition (Patel et al, 2004). Additionally, expression of E7 in an osteosarcoma cell line derivative prevented downregulation of Plk1 after arrest triggered by induction of p14ARF (Jackson et al., 2005). Altogether, these analyses highlight subtle differences in Plk1 regulation depending on the arrest stimuli and the cell types utilized. With regard to HPV-induced carcinogenesis, our results examine the effects of E6 and E7 function on Plk1 in response to DNA damage in a more relevant tissue culture system, where E6 appears to have a dominant Oncogene
Upregulation of Plk1 by HPV-16 E6 results in tetraploidy A Incassati et al
2450
regulatory effect on Plk1 protein over any effect E7 may bestow. In sum, the results presented here provide evidence that Plk1 may act downstream of p53 degradation by E6 to cause tetraploidy but not necessarily contribute to the bypass of arrest caused by adriamycin exposure. In the context of HPV-mediated cellular transformation, our data suggest that the ability of cells to bypass DNA damage arrest and become tetraploid are actually two distinct phenotypes acquired during the transition of a normal cell into a cancerous one. Although both phenotypes can result from loss of p53 function, the contribution of Plk1 to the generation of tetraploidy but not bypass of arrest downstream of p53 points to a divergence in the p53 effector response to different stimuli, with Plk1 being a target of p53 in response to one branch of the effector pathway. Materials and methods Cell culture HFKs isolated from neonatal foreskins were cultured in Epilife media containing human keratinocyte growth supplement (Cascade Biologics),1% penicillin-streptomycin. FNX GP packaging cells (ATCC) were maintained in DMEM containing 10% Fetal Clone 1 serum (FC1) (HyClone), 1% sodium pyruvate, 1% penicillin–streptomycin. Adriamycin (doxorubicin–HCl) treatments were 0.1 mg/ml for 24 h, unless otherwise stated. Experiment was carried out using batches of keratinocytes from different foreskins. Retroviral constructs and infections E6/E7-expressing cells were created using the pBabe retroviral system (Morgenstern and Land, 1990). pBabe puro (pBP) backbone was provided by H. Land. E6/E7 and the E6 D123127 (Foster et al., 1994) in tandem with E7 were cloned into pBP between BamH1-EcoR1 restriction sites. The pSuper.retro (pSR) vector (OligoEngine) and the p53 siRNA molecule were described previously (Brummelkamp et al., 2002). The p53 RNAi sequence is 50 -GACTCCAGTGGTAATCTAC-30 . The scrambled RNAi sequence is 50 -ACAAACACGAGTG TGCCCA-30 . Murine pcDNA3-Plk1 construct was a gift from E Nishida. Plk1 was subcloned out of pcDNA3 using EcoR1Not1 restriction sites and inserted into the EcoR1-Sal1 sites of pBabe through sticky-blunt ligation. FNX cells were transfected by calcium phosphate method as described by the manufacturer’s protocol (Clontech) with equal concentrations of pBP or pSR vector and the vesicular stomatitis virus envelope gene. Virus was filtered through a 0.45 mm syringe after 48 h and centrifuged at 15 000 r.p.m. for 90 min at 41C. Cells were infected in media containing virus and 8 mg/ml sequabrene (Sigma) for 6 h, then selected for 2–4 days in 1.25 mg/ml puromycin. FACS analysis Cells were pulsed with 10 mM BrdU for 45 min, harvested and stored in 70% ethanol at 41C. Nuclei were isolated by resuspending cells for 20 min in 0.2 mg/ml pepsin in 2N HCl/ PBS at 371C. Nuclei were rinsed in 0.5% FC1/0.5% PBST (PBS and 0.5% Tween-20) twice and incubated in 2% FC1/ PBS for 20 min. Nuclei were treated with 10 mg/ml anti-BrdU conjugated to flourescein isothiocyanate (FITC) (Roche) for 1 h in the dark. Nuclei were washed with 0.5% FC1/0.5% PBST and resuspended in 1 mg/ml RNase A for 30 min in the Oncogene
dark. Nuclei were centrifuged and resuspended in 20 mg/ml Propidium Iodide. 15 000 events were obtained on the FACSCalibur (Elite) flow cytometer and data was analysed using Multicycle software. RT-PCR analysis Cells pellets were stored at 801C. Total RNA was isolated using the RNeasy Kit according to the manufacturer’s instructions (Qiagen). RNA, 2 mg, were reverse transcribed in 25 ml volumes containing 400 mM dNTPs, 10 ng random hexamers (Integrated DNA Technologies (IDT)), RNaseOut, 5 U of AMV RT and buffer (Promega). Samples were incubated at 421C for 1 h. PCR reactions were performed using the following primer sets: HPV-16 E6 forward 50 -GCGGATCCATGTTTCAGG-30 and E7 reverse 50 -GCGAATTCTTATGGTTTCTGAG-30 ; HPV-6 E6 forward 50 -GCGGATCCATTATGGAAAGTG CA-30 and E7 reverse 50 -GCGAATTCTTATGTCTTCGG30 ; HPV-11 E6 forward 50 -GCGGATCCATTATGGAAAGTAAA-30 and E7 reverse 50 -GCGAATTCTTATGGTTTTG G-30 ; and glyceraldehyde-3-phosphate dehydrogenase forward 50 -CCCCTCTGCTGATGCCCCCATGTT-30 and reverse 50 -GAGCTTCCCGTCTAGCTCAGGGAT-30 . The cDNAs were diluted 1:5 and used in 25 ml PCR reactions containing PCR buffer (Promega), 1.5 mM MgCl2 (Promega), 20 mM primers (IDT), 400 mM dNTPs, 1 U Taq polymerase (Promega). Samples were amplified by PCR at 951C for 1 min, then 25 cycles of 951C for 15 s, 521C for 30 s, and 721C for 45 s. Western blotting Cell pellets were lysed in radioimmunoprecipitation assay buffer (1 PBS, 1% NP40, 0.5% sodium deoxycholate, 0.1% sodium dodecylsulfate, 1:100 protease inhibitor cocktail (Sigma)) on ice, centrifuged, and quantitated in triplicate by Bradford assay. Protein, 50 mg, was aliquoted into sample buffer and run on SDS–polyacrylamide gel. Protein was transferred to 0.2 mm nitrocellulose membrane. Ponceau S stain (Sigma) was used to verify equal loading. Membranes were blocked overnight at 41C in 5% milk/PBST. Primary antibodies were used as follows: anti-Polo-like kinase-1 cocktail (Zymed) 1:2000, anti-p53 (BD PharMingen) 1:2000, antiphospho-p53 ser-15 (Cell Signaling) 1:2000, HPV-16 E7 (Zymed) 1:250 for 1 h in milk/PBST. Membranes were washed in 0.2% PBST and incubated with goat anti-mouse IgG-HRP (Santa Cruz) 1:2000 in milk/0.2% PBST for 1 h. Membranes were washed in 0.2% PBST and visualized with Western lightning chemiluminesence reagent (Perkin Elmer) or SuperSignal West Femto Max Sensitivity reagent (Pierce) on a ChemiImager 5500 (Alpha Innotech Corp.). Luciferase assays and transient transfections The pGL2-Plk promoter luciferase reporter was a gift from T Uchiumi. The pBS-U6 siRNA plasmid was a gift from Y Shi. The siRNA sequence in the pBS-U6 plasmid targeting p53 is 50 -GGACAGCCAAGTCTGTGACT-30 , targeting the firefly luciferase gene is 50 -GGACAAGACAATTGCACTGA-30 and targeting Akt3 is 50 -GGCAAGAAGAGGAGAGAATGA- 30 . The pcDNA-p53 expression vector was a gift from B Vogelstein. The pSG5 vector was modified to reverse the EcoR1-BamH1 restriction sites. E6/E7 was sub-cloned into BamH1-EcoR1 sites. For luciferase assays, cells were seeded in six-well dishes 150 000 cells/well. Fugene 6 transfection reagent (Roche) was used as instructed by the manufacturer. DNA (1 mg) and Fugene (3 ml ) were diluted in Opti-MEM 1 serum-free media (Gibco) and added to wells containing 2 ml Epi-life media.
Upregulation of Plk1 by HPV-16 E6 results in tetraploidy A Incassati et al
2451 Media was refreshed 6 h post-transfection. Cells were lysed directly in wells after 48 h using 2 Reporter Lysis Buffer (Promega). Lysate was mixed with Luciferase Assay Reagent (Promega) following manufacturer protocol and arbitrary luciferase units were assessed by a LumiCount Luminometer (Packard). For other transient transfections, the protocol above was followed except that 10 mg DNA, 30 ml Fugene 6, and 6 ml media were used per 10 cm plate.
Acknowledgements We sincerely thank Dirk Bohmann, Jiyong Zhao, and McCance lab members for critical reading of the manuscript. This work was funded by NIH Grant NIDCR DE13526 to DJ M. AI was supported by the Rochester Training Program in Oral Infectious Diseases (T32-DE07165).
References Brummelkamp TR, Bernards R, Agami R. (2002). Science 296: 550–553. Burd EM. (2003). Clin Microbiol Rev 16: 1–17. Cahill DP, Lengauer C, Yu J, Riggins GJ, Willson JK, Markowitz SD et al. (1998). Nature 392: 300–303. Cross SM, Sanchez CA, Morgan CA, Schimke MK, Ramel S, Idzerda RL et al. (1995). Science 267: 1353–1356. Demers GW, Espling E, Harry JB, Etscheid BG, Galloway DA. (1996). J Virol 70: 6862–6869. Dyson N, Howley PM, Munger K, Harlow E. (1989). Science 243: 934–937. Eichten A, Westfall M, Pietenpol JA, Munger K. (2002). Virology 295: 74–85. Flatt PM, Price JO, Shaw A, Pietenpol JA. (1998). Cell Growth Differ 9: 535–543. Foster SA, Demers GW, Etscheid BG, Galloway DA. (1994). J Virol 68: 5698–5705. Galipeau PC, Cowan DS, Sanchez CA, Barrett MT, Emond MJ, Levine DS et al. (1996). Proc Natl Acad Sci USA 93: 7081–7084. Guan R, Tapang P, Leverson JD, Albert D, Giranda VL, Luo Y. (2005). Cancer Res 65: 2698–2704. Hanks S, Coleman K, Reid S, Plaja A, Firth H, Fitzpatrick D et al. (2004). Nat Genet 36: 1159–1161. Hawley-Nelson P, Vousden KH, Hubbert NL, Lowy DR, Schiller JT. (1989). EMBO J 8: 3905–3910. Jackson MW, Agarwal MK, Yang J, Bruss P, Uchiumi T, Agarwal ML et al. (2005). J Cell Sci 118: 1821–1832. Kho PS, Wang Z, Zhuang L, Li Y, Chew JL, Ng HH et al. (2004). J Biol Chem 279: 21183–21192. Knecht R, Elez R, Oechler M, Solbach C, von Ilberg C, Strebhardt K. (1999). Cancer Res 59: 2794–2797. Lanni JS, Jacks T. (1998). Mol Cell Biol 18: 1055–1064. McMurray HR, McCance DJ. (2004). J Virol 78: 5698–5706.
Meraldi P, Honda R, Nigg EA. (2002). EMBO J 21: 483–492. Michel L, Diaz-Rodriguez E, Narayan G, Hernando E, Murty VV, Benezra R. (2004). Proc Natl Acad Sci USA 101: 4459–4464. Morgenstern JP, Land H. (1990). Nucleic Acids Res 18: 3587–3596. Mundt KE, Golsteyn RM, Lane HA, Nigg EA. (1997). Biochem Biophys Res Commun 239: 377–385. Munger K, Phelps WC, Bubb V, Howley PM, Schlegel R. (1989). J Virol 63: 4417–4421. Nigg EA. (2001). Nat Rev Mol Cell Biol 2: 21–32. Patel D, Incassati A, Wang N, McCance DJ. (2004). Cancer Res 64: 1299–1306. Scheffner M, Werness BA, Huibregtse JM, Levine AJ, Howley PM. (1990). Cell 63: 1129–1136. Smits VA, Klompmaker R, Arnaud L, Rijksen G, Nigg EA, Medema RH. (2000). Nat Cell Biol 2: 672–676. Southern SA, Lewis MH, Herrington CS. (2004). Br J Cancer 90: 1949–1954. Spankuch-Schmitt B, Bereiter-Hahn J, Kaufmann M, Strebhardt K. (2002). J Natl Cancer Inst 94: 1863–1877. Storchova Z, Pellman D. (2004). Nat Rev Mol Cell Biol 5: 45–54. Thomas JT, Laimins LA. (1998). J Virol 72: 1131–1137. Thompson DA, Belinsky G, Chang TH, Jones DL, Schlegel R, Munger K. (1997). Oncogene 15: 3025–3035. Weichert W, Denkert C, Schmidt M, Gekeler V, Wolf G, Kobel M et al. (2004). Br J Cancer 90: 815–821. Weichert W, Kristiansen G, Winzer KJ, Schmidt M, Gekeler V, Noske A et al. (2005). Virchows Arch 446: 442–450. Werness BA, Levine AJ, Howley PM. (1990). Science 248: 76–79. zur Hausen H. (2000). J Natl Cancer Inst 92: 690–698.
Oncogene
