The tumour suppressor p53 negatively controls cell cycle progression in response to perturbed ribosome biogenesis in mammalian cells, thus coordinating growth with proliferation. Unlike mammalian cells, p53 is not involved in the growth control of proliferation in yeasts and flies. We investigated whether a p53-independent mechanism of response to inadequate ribosome biogenesis rate is also present in mammalian cells. We studied the effect of specific inhibition of rRNA synthesis on cell cycle progression in human cancer cell lines using the small-interfering RNA procedure to silence the POLR1A gene, which encodes the catalytic subunit of RNA polymerase I. We found that interference of POLR1A inhibited the synthesis of rRNA and hindered cell cycle progression in cells with inactivated p53, as a consequence of downregulation of the transcription factor E2F-1. Downregulation of E2F-1 was due to release of the ribosomal protein L11, which inactivated the E2F-1-stabilising function of the E3 ubiquitin protein ligase MDM2. These results demonstrated the existence of a p53-independent mechanism that links cell growth to cell proliferation in mammalian cells, and suggested that selective targeting of the RNA polymerase I transcription machinery might be advisable to hinder proliferation of p53-deficient cancer cells.
A proliferating cell must increase its own structural and functional components before division to ensure normal-sized daughter cells (Thomas, 2000). In cells stimulated to proliferate, the demand for an increased rate of protein synthesis is accomplished by an increased ribosome biogenesis, and the rate of ribosome biogenesis controls the cell cycle progression (Volarević et al., 2000; Derenzini et al., 2005). The derangement of cell growth caused by a defective ribosome biogenesis induces a checkpoint control that prevents the G1–S phase transition. Similarly to the mechanism by which an aberrant ribosome biogenesis activates the checkpoint, several studies pointed to the tumour suppressor p53 as the key factor in the response to ribosome biogenesis perturbations in mammalian cells (Mayer and Grummt, 2005; Opferman and Zambetti, 2006; Panic et al., 2007). Stabilised p53 activates the transcription of p21WAF1, an inhibitor of the cyclin-dependent kinases, which, by hindering pRb phosphorylation, prevents the cell from overcoming the G1–S phase restriction point (Sherr and Roberts, 1999). The central role of p53 in the relationship between growth and proliferation appears to be reached during evolution from simple eukaryotes to metazoans. In Saccharomyces cerevisiae, p53 is not present. Instead, Whi5, the yeast functional equivalent to the human tumour suppressor pRb, controls the mechanism that links rRNA synthesis to cell proliferation (Bernstein et al., 2007). In Drosophila melanogaster, a fly homologue of p53 is present, but it has no function in halting cell proliferation that results from perturbed growth (Grewal et al., 2007).
Here, we wondered whether p53-independent mechanisms controlling the cross-talk between growth and proliferation are also present in mammalian cells. Indeed, some doubt was cast on the exclusive role of p53 in determining cell cycle arrest after rRNA inhibition. The inhibition of the rRNA transcription by actinomycin D did not induce any change in cell cycle progression, but only in human tumour cells in which both p53 and pRb expression was silenced (Montanaro et al., 2007). This indicated the co-existence of a p53-independent, pRb-mediated mechanism of control of cell cycle progression through changes in the rRNA synthesis rate. Even though the dose of actinomycin D used in the latter study is generally accepted as specifically inhibiting the RNA polymerase I transcription, because this drug acts as a DNA-intercalating agent, the possibility that alterations in cell cycle progression were not exclusively due to inhibition of rRNA synthesis cannot be ruled out. Therefore, to study the relationship between the rRNA synthesis and cell cycle progression (and between cell growth and cell proliferation), we first set up an experimental approach for the selective inhibition of rRNA transcription. This entailed the use of the small interfering (si)RNA procedure to silence the POLR1A gene, which encodes the catalytic subunit of the RNA polymerase I, the enzymatic complex exclusively engaged in rRNA transcription (Grummt, 2003).
We found that the silencing of POLR1A induced a progressive reduction of the synthesis of rRNA in human cancer cell lines without affecting the RNA-polymerase-II- and -III-dependent transcription of ribosome components. Therefore, we used POLR1A silencing to study the significance of p53 and pRb in response to the inhibition of rRNA synthesis. In addition to the expected activation of the p53–p21–pRb inhibitory pathway in cells with functional p53, we observed that POLR1A silencing also hindered cell cycle progression at the G1–S phase transition in human cancer cell lines with inactivated p53. We observed that in these cells, RB1 silencing removed cell cycle progression inhibition, and so we investigated the mechanism involved in the pRb-mediated control of cell cycle progression. We demonstrated that silencing of POLR1A causes downregulation of E2F-1, which is mediated by the ribosomal protein rpL11. These reduced levels of E2F-1 slow down cell cycle progression in p53-deficient, but pRb-proficient, cells.
POLR1A silencing downregulates the synthesis of rRNA
To silence the POLR1A gene, we transfected the p53- and pRb-proficient U2OS and HCT-116 human cancer cell lines with a cocktail of three different siRNAs, and evaluated the efficacy of the RNA interference (RNAi) procedure by measuring the level of the POLR1A mRNA using real-time RT-PCR analysis. POLR1A mRNA was greatly reduced in U2OS and HCT-116 RNAi cells compared with levels in cells transfected with control sequences, as early as 24 hours after the end of the silencing procedure, and they remained low for at least 72 hours (Fig. 1a). The reduction of POLR1A expression was accompanied by a strong inhibition of rRNA synthesis, as ascertained by using two different approaches: (1) quantitative real-time RT-PCR analysis of the 45S pre-rRNA expression (Fig. 1b), and (2) immunofluorescence detection of nucleolar 5-fluorouridine incorporation into nascent rRNA (Fig. 1c). Unlike the effect on rRNA synthesis, POLR1A siRNA did not reduce the mRNA expression of ribosomal proteins L11 and L5 (transcribed by RNA polymerase II) or the expression of 5S rRNA (transcribed by RNA polymerase III) (Fig. 1d), evaluated 48 hours after the end of the silencing procedure. Furthermore, we found that at 48 hours, POLR1A siRNA did not significantly reduce the total protein synthesis level (Fig. 1e). Previously, it has been reported that impairment of ribosome biogenesis by depletion of ribosomal protein S6 for 24 hours led to a reduction of global translation in lung adenocarcinoma A549 cells (Fumagalli et al., 2009). For this reason, to substantiate our data on the effect of POLR1A siRNA on protein synthesis, we wondered whether POLR1A silencing induced a change in the total amount of ribosomes. Densitometry analysis of the amount of 28S and 18S rRNA showed no difference between control and POLR1A-silenced U2OS cells after 48 hours (supplementary material Fig. S1). Moreover, in a recent study, we observed that POLR1A siRNA did not induce variations in ribosomal protein L5 and L11 synthesis 48 hours after the end of the silencing procedure (Donati et al., 2011). The early inhibition of protein synthesis observed by Fumagalli and colleagues (Fumagalli et al., 2009) might be due to the specific approach to hinder ribosome biogenesis. The ribosomal protein S6 is used stoichiometrically in building up new ribosomes: therefore, protein S6 depletion might induce a rapid downregulation of ribosome production, which was not found to occur after POLR1A interference. The fact that we did not find a significant quantitative change of ribosome content in POLR1A-silenced cells in comparison with control cells can be explained by considering that after POLR1A RNAi, progressive reduction, and not an arrest of rRNA synthesis occurred (see Fig. 1c). This probably allowed cells that were distributed throughout the G1 phase to accumulate in late G1 phase and cells that were in S phase to accumulate in G2 phase. In other words, after POLR1A silencing, the cell population was mainly in late G1 and G2 phase, when cells are characterised by a high complement of ribosomes.
Silencing of POLR1A hinders cell proliferation by activating the p53–p21–pRb inhibitory pathway
Because these results showed that POLR1A interference selectively reduced the synthesis of rRNA, we wanted to determine whether this way of inhibiting rRNA synthesis might have the same effect on p53 stabilisation, pRb phosphorylation, cell cycle progression and cell proliferation, in terms of perturbation of ribosome biogenesis as observed for other reported methods. Western blot analysis performed 48 hours after the end of the transfection procedure showed an accumulation of p53 in POLR1A-silenced U2OS cells, an increased expression of p21 and the disappearance of the phosphorylated form of pRb (Fig. 2a). The same results were also obtained with the HCT-116 cell line (data not shown). To evaluate the effect of the activation of the p53–p21–pRb pathway on cell proliferation, we first studied the changes induced by POLR1A siRNA on the proliferation rate of U2OS and HCT-116 cells, and of MCF-7 cells, which are another p53- and pRb-proficient cancer cell line. As expected, in the three cell types, the reduction of ribosome biogenesis hindered the expansion of cell populations, as evaluated with the Crystal Violet assay (Carnero et al., 2000), 5 days after the silencing procedure (Fig. 2b). We also evaluated the effect of POLR1A RNAi on the number of DNA-synthesising cells, by measuring the bromodeoxyuridine-labelling index (BrdU-LI: percentage of cells incorporating BrdU into DNA) of U2OS cells (Fig. 2c). We observed a significant reduction in the number of BrdU-labelled cells 48 hours after the silencing procedure. Finally, we measured the modification induced by POLR1A RNAi on cell cycle phase distribution by flow-cytometry analysis of the DNA content in the U2OS cells. An accumulation of cells in G1 phase and a concomitant reduction of S-phase cells was observed 48 hours after POLR1A silencing (Fig. 2d).
Silencing of POLR1A hinders cell cycle progression in a pRb-dependent manner in cells lacking functional p53
Our results demonstrated that POLR1A RNAi is a reliable method for specifically inhibiting rRNA synthesis, which activates the established p53-mediated control of cell proliferation (Mayer and Grummt, 2005). For this reason, we used POLR1A silencing to define the actual role of p53 and pRb in the control of cell proliferation after rRNA synthesis inhibition. Regarding p53, we silenced POLR1A expression in U2OS and HCT-116 cells, which were also silenced for TP53 (the gene encoding p53) expression. Simultaneous silencing of POLR1A and TP53 greatly reduced the expression of both POLR1A and TP53 mRNA in U2OS cells, 48 hours after the end of the silencing procedure (Fig. 3a). Western blot analysis showed the absence of p53 expression in TP53-silenced and in TP53- and POLR1A-silenced cells (Fig. 3b; see also immunocytochemical data, supplementary material Fig. S2a). Similar results were obtained using HCT-116 cells. As expected, the proliferation rate evaluated both with the Crystal Violet assay (Fig. 3c) and by immunocytochemical analysis of the BrdU-LI (Fig. 3d), was greatly reduced in p53-proficient cells after POLR1A RNAi. However, we observed that after POLR1A RNAi, a significant reduction of BrdU-LI also occurred in TP53-silenced cells, albeit at a lower level than that in control cells. To establish the role of pRb in response to inhibition of ribosome biogenesis in cells lacking the p53 function, it would have been necessary to perform a triple silencing procedure involving RB1, TP53 and POLR1A. Because a triple interference procedure did not allow a satisfactory silencing of the three genes and caused a high mortality rate in the cells (data not shown), we inactivated p53 function in HCT-116 cells by viral transduction of a retroviral vector expressing p53DD, a dominant-negative, truncated form of murine p53 (Shaulian et al., 1995). In these cells (p53DD) p53 significantly accumulated in the nuclei and was functionally inactive (Derenzini et al., 2009). In both p53DD cells and the relevant control cells transduced with the empty retroviral vector (pBABE), the interference of both RB1 and POLR1A heavily reduced the expression of the respective mRNAs (Fig. 4a). The immunocytochemical analysis confirmed the absence of pRb in the RB1-silenced cells (supplementary material Fig. S2b). In pBABE cells, POLR1Asilencing stabilised p53, induced p21 expression and reduced the phosphorylated form of pRb, as revealed by the use of an antibody against S608-phosphorylated pRb (Fig. 4b). In p53DD cells, p53 appeared to be equally accumulated in controls and in POLR1A-silenced cells, whereas neither p21 expression nor the disappearance of the phosphorylated form of pRb was observed after POLR1A silencing, thus demonstrating that p53 was inactive in these cells (Fig. 4b). The reduction of the expression of the phosphorylated form of pRb after POLR1A silencing in pBABE, but not in p53DD cells, was also detected after treatment with antibody against all forms of pRb (supplementary material Fig. S3). We then measured the changes of BrdU-LI in pBABE and p53DD HCT-116 cells after POLR1A RNAi, according to the expression of RB1.
We observed that RB1 silencing per se had a moderate inhibitory effect on BrdU incorporation in both pBABE and p53DD cells (although it was stronger in the former). This might be explained by the fact that depletion of pRb induces accumulation of DNA double-strand breaks (Pickering and Kowalik, 2006; Derenzini et al., 2008). POLR1A RNAi induced a reduction of the BrdU-LI in pBABE and to a lesser extent in p53DD cells, but not in cells silenced for RB1 expression (Fig. 4c). The cytofluorimetric analysis of the cell population distribution within cell cycle phases indicated that POLR1A RNAi induced an accumulation in the G1 phase and a reduction in the S phase, which was higher in pBABE than in p53DD cells (Fig. 4d). These changes were no longer observed after POLR1A RNAi in cells silenced for RB1 expression, which is also in agreement with the results obtained from the analysis of the BrdU-LI.
Silencing of POLR1A downregulates expression of E2F-1
Regarding the mechanism by which pRb might control cell cycle progression in the absence of the p53 function, we focused our attention on the expression of E2F-1, the transcription factor whose activity is negatively controlled by pRb. Therefore, we measured the expression of E2F-1 protein in U2OS, MCF-7 and in pBABE and p53DD HCT-116 cells, after POLR1A silencing. E2F-1 protein expression was reduced in the four cell lines 48 hours after the end of the silencing procedure (Fig. 5a). Considering that E2F-1 is mainly expressed during the S phase of the cell cycle, we wondered whether the reduction in the E2F-1 protein observed in cells after POLR1A RNAi might be a consequence of the changes induced in the cell cycle progression, i.e. cell accumulation in the G1 phase. Therefore, we evaluated the changes of E2F-1 expression after POLR1A RNAi in pBABE and p53DD cells silenced for RB1 expression, in which the inhibition of rRNA synthesis did not modify cell cycle progression. Western blot analysis showed that POLR1A RNAi also reduced the expression of E2F-1 in these cells (Fig. 5b). The same occurred in MDA-MB-468 cells, which lack both pRb and p53 functions (Wang et al., 1993), and in which the BrdU-LI was quite similar to that of cells transfected with control sequences (Fig. 5c). This indicates that the downregulation of E2F-1 expression in POLR1A-silenced cells was not the consequence of changes in cell cycle progression. Because E2F-3, another member of the E2F family, is involved in the control of cell cycle progression to S phase (Leone et al., 1998), and therefore can influence the cell proliferation rate, we evaluated the effect of POLR1A RNAi on the expression of E2F-3 in U2OS and HCT-116 cells. Forty-eight hours after the end of the silencing procedure, E2F-3 expression was not modified in comparison with control cells (supplementary material Fig. S4).
To investigate whether the reduction of E2F-1 protein expression caused by rRNA synthesis inhibition might be responsible for the hindering of cell cycle progression in p53-deficient cells, we carried out a silencing procedure of the E2F1 gene. This led to a downregulation of E2F1 mRNA expression to about 60% of the control level, in p53DD HCT-116 cells either silenced or not silenced for RB1 expression (Fig. 5d). Forty-eight hours after the end of the E2F1 silencing procedure, cells showed a reduction of E2F-1 protein expression of about 50% of the protein level of control cells (Fig. 5e), similarly to that observed after POLR1A RNAi. We then measured the changes of BrdU-LI in p53DD HCT-116 cells after E2F1 RNAi, according to the expression of RB1. E2F1 RNAi induced a reduction of the BrdU-LI in p53DD cells, but not in cells silenced for RB1 expression (Fig. 5f). Therefore, these data indicated that the downregulation of E2F-1 protein expression induced by POLR1A silencing was actually responsible for the reduced cell proliferation rate in cells with inactivated p53, and that this inhibition was conditioned by RB1 expression. Consistent with these results was the observation that E2F-1 overexpression induced by transfection of an expression plasmid that encodes the E2F-1 protein (Dick and Dyson, 2003) into p53DD cells (Fig. 5g), overwhelmed the reduction in proliferation rate caused by POLR1A silencing, as evaluated by BrdU-LI (Fig. 5h).
At this point, we tried to clarify how the reduction of E2F-1 expression hindered cell proliferation in p53-deficient cells. Thus we evaluated the transcription level of four E2F-1 target genes (CCNE1, RRM2, MCM7, TYMS) associated with cell cycle progression. We observed that POLR1A silencing induced a reduction in the transcription of the E2F-1 target genes in p53DD cells 48 hours after the end of the silencing procedure, whereas it did not modify their transcription in cells also silenced for RB1 expression (Fig. 5i). This strongly suggested that the reduction in the products of genes necessary for cell cycle progression was the cause of the hindered cell proliferation in cells lacking functional p53 after POLR1A silencing. The finding that this inhibitory effect was rescued by RB1 silencing suggested that a sufficient amount of E2F-1 protein was left to promote cell cycle progression after inhibition of rRNA synthesis. In the pRb-proficient cells, this amount of E2F-1 should be mainly bound to pRb, and therefore would be functionally inactive.
To clarify the latter point, we measured the amount of the free E2F-1 protein in p53DD cells silenced for POLR1A expression, because in these cells pRb is not activated by dephosphorylation in response to the inhibition of ribosome biogenesis. After immunoprecipitation of pRb, the amount of free (not linked to pRb) E2F-1, proved much lower in POLR1A-silenced cells compared with control cells (Fig. 5j). Therefore, the hindered cell cycle progression registered in pRb-proficient p53-deficient cells after POLR1A silencing, is due to a reduced availability of free E2F-1.
POLR1A silencing increases proteasomal degradation of E2F-1
To obtain a mechanistic insight into how POLR1A silencing downregulates E2F-1 protein expression, we first investigated at which level – transcriptional or post-transcriptional – the cause of the reduction in E2F-1 expression could be found. Time-course analysis revealed that E2F-1 protein expression was reduced in HCT-116 cells as early as 24 hours after the end of the POLR1A silencing procedure. No change in the expression of E2F1 mRNA occurred at that time, indicating a post-transcriptional mechanism for the downregulation of the E2F-1 protein (Fig. 6a). We found that E2F1 mRNA expression was reduced 48 hours after POLR1A silencing, which might be explained by the fact that the E2F-1 protein controls its own gene expression (Hsiao et al., 1994). We also evaluated whether the reduction of E2F-1 expression after POLR1A silencing was due to a reduced efficiency of E2F1 mRNA translation. We found that the level of the E2F1 mRNA associated with polyribosome fractions was similar both in control and in POLR1A-silenced U2OS cells, as evaluated by real-time RT-PCR (supplementary material Fig. S5), thus indicating that POLR1A interference did not induce variations in E2F-1 protein synthesis. Because there is evidence that the E2F-1 protein level is controlled by ubiquitin-dependent proteolysis, which is regulated by the binding to the ubiquitin-protein ligase SCFSKP2 during the S–G2 phase (Marti et al., 1999), we considered the possibility that the reduction of the E2F-1 protein expression after POLRIA silencing might be due to its increased degradation by the ubiquitin-protein ligase SCFSKP2 pathway. Following SKP2 silencing for 48 hours, the expression of the SKP2 mRNA was highly reduced both in control cells and POLR1A-silenced HCT-116 cells (after silencing for 48 hours, the POLR1A mRNA was reduced both in control and SKP2-silenced cells) (supplementary material Fig. S6). E2F-1 protein was increased in both control and POLR1A-silenced HCT-116 cells, in comparison with cells not silenced for SKP2 expression. However, the quantitative difference in E2F-1 expression between POLRIA-silenced and control cells was maintained (Fig. 6b), thus demonstrating the presence of an SKP2-independent mechanism of E2F-1 degradation after inhibition of rRNA synthesis. However, the unspecific block of proteasome activity with the small inhibitor molecule MG-132 completely reversed the reduction of E2F-1 protein in POLR1A-silenced cells (Fig. 6c). Therefore, these results demonstrated that E2F-1 protein expression is reduced by increased proteasome degradation independently of the SCFSKP2 pathway.
The oncoprotein MDM2 binds to the E2F-1 protein and protects the factor from proteasome-dependent degradation (Zhang et al., 2005), whereas several ribosomal proteins have been shown to inhibit the MDM2 function after the perturbation of ribosome biogenesis (Zhang and Lu, 2009). In a series of experiments carried out with the aim of clarifying the mechanisms involved in p53 stabilisation after inhibition of rRNA synthesis, we observed that POLR1A RNAi allowed ribosomal proteins that were no longer used for ribosome biogenesis to bind at greater levels to MDM2, thus increasing p53 stabilisation consequent to the reduced MDM2-mediated p53 proteasomal degradation (results not shown). We wondered whether the reduction of the E2F-1 protein expression might be due to its accelerated degradation as a consequence of the ribosomal protein interaction with MDM2. Therefore, we evaluated the importance of the ribosome protein L11 – a major regulator of MDM2 activity (Zhang and Lu, 2009) – in the mechanism leading to the downregulation of E2F-1 protein expression. For this purpose, we evaluated the effect of RPL11 silencing on the expression of E2F-1 protein in pBABE and p53DD cells either silenced or not silenced for POLR1A expression. Double RPL11 and POLR1A RNAi strongly reduced the expression of the relative mRNAs, whereas RPL11 RNAi alone increased POLR1A expression (Fig. 6d), very likely as a consequence of Myc transcriptional activation (Dai et al., 2007). Western blot analysis showed that after RPL11 silencing, the level of E2F-1 expression did not decrease in cells silenced for POLR1A compared with control rpL11-proficient cells (Fig. 6e), thus demonstrating that rpL11 availability is necessary for the downregulation of E2F-1. At this point, to clarify the role played by MDM2 in the downregulation of E2F-1 protein expression after inhibition of rRNA synthesis, we transfected p53DD cells with an expression plasmid encoding MDM2 (Marston et al., 1994) and then evaluated the effect of the POLR1A RNAi on the expression of E2F-1 protein. We found that the MDM2 overexpression was associated with an increased amount of E2F-1 protein, and that POLR1A RNAi did not reduce the expression of the E2F-1 protein in cells harbouring excess MDM2 (Fig. 6f). Furthermore, the proliferation rates of cells overexpressing MDM2, either silenced or not silenced for POLR1A expression, were not significantly different from each other (Fig. 6g). These results were consistent with a mechanism of E2F-1 protein downregulation involving the release of the ribosomal protein L11 after the inhibition of rRNA synthesis, which leads to MDM2 inactivation that no longer preserves E2F-1 from proteasome degradation. Immunoprecipitation experiments using anti MDM2 antibodies demonstrated that this was in fact the case. Western blot analysis of E2F-1 expression showed that in HCT-116 cells silenced for POLR1A expression, the amount of E2F-1 immunoprecipitated with MDM2 was reduced in comparison with control cells and with cells silenced for RPL11 expression (Fig. 6h). However, the amount of rpL11 co-precipitated with MDM2 was greater in POLR1A-silenced than in control HCT-116 cells.
In the present study, by silencing the gene that codes for the catalytic subunit of RNA polymerase I, we had the chance to evaluate the effect on cell cycle kinetics of rRNA synthesis inhibition alone. Previous studies were conducted using models based on: (1) induction of a defective nucleolar protein necessary for both rRNA processing and the 60S ribosome subunit production (Pestov et al., 2001; Hölzel et al., 2005); (2) conditional deletion of the gene encoding the 40S ribosome protein S6 (Sulić et al., 2005); and (3) genetic inactivation of TIF-IA (Yuan et al., 2005), which is an essential RNA polymerase I transcription factor that also regulates the levels of both 5S rRNA and mRNAs encoding ribosomal proteins in Drosophila (Grewal et al., 2007). In these cases, the effect on cell proliferation could be ascribed not only to the inhibition of rRNA transcription, but also to ribosome biogenesis errors, decreased number of ribosomes, changes in the ratio of the 40S and 60S ribosome subunits, and defective ribosome translation (Panic et al., 2007). In agreement with what has been demonstrated to occur with the use of these models of deranged ribosome biogenesis, we found that POLR1A silencing hindered cell proliferation in p53- and pRb-proficient human cancer cell lines by activating the p53–p21–pRb pathway (Mayer and Grummt, 2005; Opferman and Zambetti, 2006; Panic et al., 2007). Therefore, the present results confirm that p53 stabilisation is the major mechanism by which an altered ribosomal biogenesis induces the arrest of cell cycle progression. However, our data also indicated that p53 was not the only factor controlling the relationship between ribosome biogenesis and cell proliferation. We demonstrated that the inhibition of rRNA transcription also hindered cell cycle progression at the G1–S phase transition in cells with inactivated p53. In cells with inactivated p53, silencing of RB1 rescued inhibition of cell cycle progression, thus indicating a p53-independent role for pRb in the relationship between cell growth and proliferation. We found that the downregulation of E2F-1 expression was at the basis of this p53-independent role of pRb. E2F-1 belongs to a family of transcriptional regulators called the E2Fs, which control the expression of genes whose products are important for the entry and passage throughout the S phase. In resting cells, hypophosphorylated pRb binds E2F-1, thus blocking its transactivating function on E2F target genes (Weinberg, 1995). When the cell enters the cell cycle, phosphorylation of pRb by cyclin-dependent protein kinases (Harbour et al., 1999) leaves E2F-1 free to activate the target genes involved in the synthesis of DNA. There is evidence that the expression of E2F-1 protein, similarly to other cell cycle regulatory proteins, is closely related to the cell cycle phases, i.e. it progressively accumulates during the late G1 phase, reaching its highest value during S phase and is rapidly degraded in S–G2 phase by ubiquitylation and proteasomal digestion (Marti et al., 1999). The reduction of E2F-1 expression after the inhibition of rRNA synthesis was observed in all the cell lines examined. It was independent of p53 and pRb function, it was not due to changes in the cell cycle progression, and it was sufficient to hinder cell proliferation. The latter statement was supported by the observation that downregulation of E2F-1 protein expression induced by E2F1 gene silencing, quantitatively quite similar to that observed after POLR1A-silencing, also reduced cell cycle progression. This does not exclude the idea that other factors controlled by pRb might also have a role in the control of cell proliferation after rRNA inhibition in p53-deficient cells.
Regarding the mechanism involved in the reduction of E2F-1 expression, our results excluded a control at the transcriptional level. Time-course analysis of the E2F-1 protein and mRNA expression indicated that the reduction of the protein preceded that of the mRNA, thus suggesting either a translational or post-translational control. We found that the reduced expression of E2F-1 was due to its increased degradation. The degradation of the E2F-1 protein was hindered by its binding with MDM2, which has stabilising, p53-independent properties on the protein, by inhibiting E2F-1 ubiquitylation (Zhang et al., 2005). A widely accepted mechanism of p53 checkpoint activation after alteration of ribosome biogenesis envisages that, whatever the cause of the perturbation of the ribosome biogenesis homeostasis, a nucleolar change occurs with a leakage of some ribosomal proteins, including L5, L11, L23 and S7. These proteins bind to MDM2, thus inhibiting MDM2-mediated p53 degradation. (Lohrum et al., 2003; Zhang et al., 2003; Dai and Lu, 2004; Dai et al., 2004; Jin et al., 2004; Chen et al., 2007; Zhang and Lu, 2009). We demonstrated that the same mechanism involved in p53 stabilisation was responsible for the reduction of E2F-1 expression after rRNA synthesis inhibition. We showed that, in POLR1A-silenced cells, downregulation of the E2F-1 protein was reverted by blocking the proteasomal degradation activity. In addition, overexpression of MDM2 prevented rpL11-dependent downregulation of E2F-1 protein. The observation that RB1 silencing completely rescued the cell cycle progression rate in cells silenced for POLR1A expression, indicated a key role of pRb in the mechanism by which the reduction of E2F-1 expression hindered cell proliferation in cells lacking functional p53. In pRb-proficient cells with inactivated p53, the downregulation of the E2F-1 protein after POLR1A silencing caused a reduction in the expression of the E2F-1 target genes that are necessary for the entry and progression through the S phase. RB1 silencing rescued the expression of these genes completely, and prevented the effect of the rRNA synthesis inhibition on cell proliferation. This showed that, in the absence of pRb, the reduced amount of E2F-1 was still able to sustain cell cycle progression of POLR1A-silenced cells.
The effect of POLR1A RNAi on the reduction of E2F-1 expression appeared to be a specific consequence of the inhibition of rRNA synthesis. Similar results were obtained using each of the three siRNAs making up the cocktail used for POLR1A silencing (supplementary material Fig. S7). However, the inhibition of rRNA synthesis is not always associated with a downregulation of E2F-1: actinomycin D, cisplatin and etoposide – all drugs that inhibit rRNA transcription – caused an accumulation of E2F-1 protein (Lin et al., 2001; Stevens et al., 2003). It is worth noting that these drugs are DNA-damaging agents. In addition, it has been shown that the checkpoint kinase 2, phosphorylates E2F-1 in response to the DNA-damaging agents while increasing its half-life and transcriptional activity (Stevens et al., 2003). The importance of reducing the synthesis of rRNA without using DNA-damaging agents for downregulating the E2F-1 protein was demonstrated by the observation that TIF1A silencing – which is not a DNA-damaging procedure for inhibiting rRNA transcription (Yuan et al., 2005) – did reduce the expression of the E2F-1 protein, unlike treatment with actinomycin D (supplementary material Fig. S8).
In conclusion, we demonstrate that the inhibition of rRNA synthesis downregulated the expression of E2F-1, which, in cells with inactive p53, hinders cell cycle progression in a pRb dependent manner. In p53-proficient cells, the downregulation of E2F-1 expression is only of little help to p53 in the control of proliferation after perturbed growth. Nevertheless, in the absence of a functional p53, it appeared to have a key role in hindering cell proliferation after the inhibition of rRNA synthesis. The loss of p53 function very frequently characterises human cancers, which are therefore less sensitive to cytostatic and cytotoxic drug treatment (Kigawa et al., 2001; Campling and El-Deiry, 2003; Derenzini et al., 2009). The RNA polymerase I transcription machinery has been recently proposed as a new target for the treatment of cancer (Drygin et al., 2010). Our results suggest that studies on drugs targeting the RNA polymerase I transcription apparatus, to generate antineoplastic agents would be useful. This is based on their ability to selectively inhibit rRNA synthesis, without genotoxic activity, and induce a cytostatic effect in p53-deficient cancers, also by downregulating the expression of E2F-1.
Materials and Methods
Cell lines and drug treatments
U2OS osteosarcoma cells, HCT-116 colon cancer cell line, MCF-7 and MDA-MB-468 breast cancer cell lines were obtained from American Type Culture Collection and were all grown in DMEM (Euroclone) supplemented with 10% fetal bovine serum (FBS) (Euroclone), except for the MCF-7 cells, which were cultured in RPMI 1640 (Euroclone) with 10% FBS. HCT-116 pBABE and p53DD isogenic cell lines have been described previously (Derenzini et al., 2009). The proteasome inhibitor MG-132 (Calbiochem) was used at a final concentration of 10 μM.
siRNA and plasmid transfection
The day before transfection, cells were seeded in antibiotic-free growth medium. Transfections were performed with Lipofectamine RNAiMAX (Invitrogen) in Opti-MEM medium (Invitrogen) following the manufacturer's protocol. The POLR1A, TP53, RB1, TIF-1A, E2F-1 and SKP2 genes were silenced by using a pool of different siRNAs from Invitrogen (Stealth RNAi Select) targeted against different parts of the mRNA; the RPL11 gene was silenced using the siRNA sequence described previously (Barkić et al., 2009) target sequence 1; control cells were transfected with equal amounts of Stealth RNAi Negative Control from Invitrogen.
The mammalian expression plasmid CMV-HA-E2F-1 was a gift from Frederick A. Dick (Dick and Dyson, 2003); the mammalian expression plasmid for human MDM2 was a gift from Karen H. Vousden (Marston et al., 1994). The empty plasmid pcDNA3 was used to transfect control cells. Plasmids were transfected using jetPEI or jetPRIME reagents (Polyplus-transfection) according to the manufacturer's protocol.
Quantitative real-time RT-PCR
Total RNA was isolated from harvested cells using TRI Reagent (Ambion). Reverse transcription and subsequent real-time RT-PCR analysis of cDNA were performed as already described (Derenzini et al., 2009). A set of primers and TaqMan probes were purchased from Applied Biosystems (Assay on Demand) to quantify POLR1A, TP53, RB1, E2F1, CCNE1 and TYMS mRNAs, whereas primers and UPL probes (Roche) for RPL11 and RPL5 mRNA detection were chosen with the Roche online primer design tool (Universal Probe Library). Primers for SYBR Green real-time RT-PCR analysis of 5S rRNA were as described previously (Winter et al. 2000). Primers for RRM2, MCM7, TIF1A and SKP2 mRNAs were designed using the Roche online tool (all sequences available upon request).
The housekeeping gene β-glucuronidase (Applied Biosystems) was used as an internal control. For all the primers used, we applied the following cycling conditions: 50°C for 2 minutes, 95°C for 10 minutes, 35 cycles at 95°C for 15 seconds and 60°C for 1 minute.
The steady state level of 45S pre-rRNA was determined by SYBR Green real-time RT-PCR analysis using the primers described previously (Murayama et al., 2008). Low levels of DNA contamination were checked by performing a no-RT control for each sample. Cycling conditions were as follows: 50°C for 2 minutes, 95°C for 10 minutes, 35 cycles at 95°C for 30 seconds and 60°C for 1 minute.
Immunofluorescent detection of rRNA synthesis
Immunodetection of nascent rRNA was performed by incorporation of 5-fluorouridine, according to the method described (Boisvert et al., 2000). Briefly, cells growing on coverslips were incubated with 2 mM 5-fluorouridine (Sigma) for 15 minutes, then washed in cold PBS and fixed in 2% paraformaldehyde and 1% Triton X-100 for 5 minutes for subsequent immunofluorescent staining with a specific monoclonal antibody for halogenated uridine (BU-33, Sigma). Mounting and nuclei counterstaining were performed using the Pro long antifade reagent with DAPI (Molecular Probes). Samples were observed under an Axiovert 40 fluorescence microscope (Carl Zeiss).
Quantitative evaluation of the 28S and 18S rRNA transcripts
HCT116 cells were used, either silenced for TP53 expression (Stealth RNAi Select) or transfected with equal amounts of Stealth RNAi Negative Control from Invitrogen. Also used were, pBABE and p53DD HCT116 cells. Three plates were used for each different experimental conditions. Total RNA was extracted using Trizol (Invitrogen), following the manufacturer's protocol. The 28S and 18S RNA subunits were visualised by loading in a 1% agarose gel stained with ethidium bromide an equal fraction (10%) of the total quantity of the obtained RNA pooled from three plates. The intensity of the bands was evaluated with the densitometric software GelPro analyzer 3.0 (Media Cybernetics, Silver Spring, MD).
Immunocytochemical analysis of BrdU labelling
Cells seeded on glass coverslips were exposed to 20 μM BrdU (Sigma) for 2 hours, then fixed and permeabilised in PBS containing 2% paraformaldehyde and 1% Triton X-100, for 10 minutes at room temperature. Cells were treated with HCl 4N for 10 minutes and then washed with double-distilled water. For immunocytochemical staining, cells were treated with 1.5% H2O2 for 5 minutes in the dark in order to suppress endogenous peroxidase activity. Cells were incubated for 30 minutes at room temperature in PBS containing 1% bovine serum albumin (BSA) to block aspecific staining, then washed and incubated with primary anti-BrdU monoclonal antibody (85-2C8, Novocastra) diluted in PBS containing 1% BSA at 4°C in a humidified chamber. After overnight incubation, cells were washed in PBS and incubated first with a biotinylated secondary antibody (Vector Laboratories) in PBS 1% BSA for 30 minutes, and then with the streptavidin-peroxidase conjugate (Biospa) in PBS 1% BSA for 30 minutes. The streptavidin-peroxidase complex was visualised using diaminobenzydine (Sigma). Cells were counterstained in hematoxylin, dehydrated and mounted in a synthetic medium on microscope slides.
Total protein extraction, SDS-PAGE and immunoblot analysis were performed as described (Montanaro et al., 2007). Levels of target proteins were normalised to those of β-actin. The following mouse monoclonal antibodies were used: anti-pRb (1F8, Thermo Fisher Scientific), anti-phospho-S608-pRb (51B7, Thermo Fisher Scientific), anti-p53 (BP53-12, Novocastra), anti-E2F-1 (KH95), anti-MDM2 (SMP14), anti-cyclin-D1 (A-12), anti-cyclin-E (13A3) (Santa Cruz Biotechnology), anti-p21 (SX118, Dako), anti-actin (AC-74, Sigma-Aldrich). For E2F-3 detection, a rabbit polyclonal antibody was used (C-18, Santa Cruz Biotechnology). Primary antibodies were detected with horseradish-peroxidase-conjugated secondary antibodies (GE Healthcare) and using the ECL kit (GE Healthcare) or the SuperSignal West Pico kit (Thermo Fisher Scientific). The densitometric quantification of western blots was performed using GelPro analyzer 3.0 software (Media Cybernetics).
Cells were lysed in a buffer containing 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.8% NP40, 1 mM DTT, 1 mM EDTA and Complete Protease Inhibitors Cocktail (Roche Diagnostics). For each sample, 1.0 mg of proteins were pre-cleared with Protein-A/G PLUS-Agarose beads (Santa Cruz Biotechnology) at 4°C for 30 minutes and subsequently incubated overnight at 4°C with anti-pRb (C-15, Santa Cruz Biotechnology) or anti-MDM2 (H-221, Santa Cruz Biotechnology) rabbit polyclonal antibodies. The immunocomplexes were then precipitated with Protein-A/G PLUS-Agarose beads for 2 hours at 4°C. Afterwards the beads were centrifuged, washed four times in lysis buffer, and resuspended in Laemmli buffer for subsequent immunoblot analysis. After pRB immunoprecipitation, the non-precipitated fraction was further incubated for 6 hours at 4°C with anti-E2F-1 polyclonal antibodies (C-20, Santa Cruz Biotechnology). The immunocomplexes were then isolated for immunoblot analysis as described above.
Evaluation of cell proliferation rate
The cell population evaluation was performed according to the Crystal Violet method as described previously (Carnero et al., 2000). Cells were formalin-fixed overnight at 4°C, stained for 30 minutes with 0.1% Crystal Violet in a 20% methanol solution, and then washed four times in double-distilled water before resolubilisation in 10% acetic acid solution. With the cell-bound samples, Crystal Violet dye was used for subsequent spectrophotometrical quantification at 595 nm wavelength.
Flow cytometry analysis of cell cycle phases
Analysis of the DNA content of cells was performed by flow cytometry by fixing cells in 70% ethanol and processing them as described (Montanaro et al., 2007), but using propidium iodide, instead of DAPI, for staining the DNA.
Isolation of polyribosomal mRNA
Subconfluent U2OS cells were washed in PBS at 4°C. The cellular pellet was lysed in 2 volumes of 10 mM Tris-HCl (pH 7.4), 10 mM NaCl, 3 mM MgCl2 0.5% NP40 for 10 minutes at 4°C. The lysates were then centrifuged at 14,000 g for 10 minutes at 4°C and the supernatant was used for the isolation of polyribosomes. Lysates were stratified onto a 15–50% sucrose gradient in 30 mM HEPES-KOH (pH 7.5), 80 mM KCl, 1.8 mM Mg-acetate, and centrifuged at 4°C for 15 hours at 40,000 g. From gradients, 1 ml fractions were collected and their absorbance was read at 260 nm. Polyribosomal fractions were pooled and centrifuged at 100,000 g for 15 hours at 4°C. RNA was extracted from pellets using Trizol reagent.
Protein synthesis evaluation
The protein synthesis rate was evaluated by incorporation of [3H]leucine added to the growth medium at a final concentration of 9 μCi/ml. After 30 minutes of incubation, the cells were washed in cold PBS, then 10% trichloracetic acid was added to lyse the cells and precipitate the cellular proteins. The proteins were resuspended in 0.2 N KOH for quantification (Bio-Rad Protein Assay). An equal amount of proteins for each sample was loaded on a β-counter for radioactivity quantification.
Student's t-test was applied to evaluate statistical significance of differences between samples.
The authors thank S. Volarević for his helpful suggestions. The authors are also thankful to Frederick Dick and Karen Vousden, for the generous gift of the E2F-1 and MDM2 expression plasmids, respectively. This work was supported by Roberto and Cornelia Pallotti's Legacy for Cancer Research, Vanini-Cavagnino grant, and MIUR (Italian Ministry of Education, University and Research: grants for Oriented Fundamental Research).